Spice Crystal Model - 70MHz

Haha- that is funny- must be the brane disease.

I normally try to overlook the occasional typo, but that one was too funny to ignore. ;-)

--
Service to my country? Been there, Done that, and I\'ve got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

"Brane" is an archaic spelling of "brain", no doubt used for effect and to stimulate ye olde brainpan.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Do crystals have any serious nonlinearities, aside from the case of gross overdrive? That could make startup interesting.

And even fundamental crystals have gobs of un-advertised resonances, some of which can be pretty close to the promised one. You just hope that the rated-frequency resonance has a much higher Q than the others, and that your circuit doesn't favor the wrong one. Overtone oscillators are actually nice because they usually (!) have an LC tank in the loop, which zaps most of the hazards.

Maybe Jim could make his loop gain peak around 70 MHz somehow. He should at least get a full-sweep plot that shows all the resonances, rather than just a Spice model of the main one.

Jim: what was wrong about the flubbed design? Did it not oscillate, or oscillate at the wrong frequency, or what?

John

Seems a bit like teaching granny to suck eggs but these are my thoughts. The parameters could vary al lot from manufacturer to manufacturer depending on the electrode diameter quality of quartz and parallelism. I would construct a device assuming scaled from 20 MHz fundemental to

70MHz fundemental as the blank will not be contoured and will have scaled plateback. Assume maximum esr of 100R AND ENSURE THAT THE CIRCUIT ACTIVITY WILL NOT OSCILLATE spurious 60MHz to 300MHz with an esr much above 120 ohms. The problem of reactance of Co less than esr is reduced due to fundemental operation and small electrode diameter but it may pay to design in some active bridge cancellation of 2pF Co . Power level must be low due say 10uW or if you can measure a crystal raise power till frequency changes 1ppm.

I was refereing to this from Fred's previous message:

KNOWN

Wow! I'm shocked. Fred actually made a funny typo. ;-)

--
Service to my country? Been there, Done that, and I\'ve got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

In article , John Larkin wrote: [...]

Yes, there is an extra loss at the small signal end. This makes it posible to adjust the gain of an oscillator so that it will sustain oscillations, but the won't start on their own.

[...]

On the SC cut crystals, the mode 10% away is still trouble with a simple LC. If you make the Q of the LC high enough, its drift starts to matter.

Lucky for JT, he most likely has a AT cut rock.

--
--
kensmith@rahul.net   forging knowledge

"Michael A. Terrell"

wrote:

try to overlook the occasional typo, but that one was too

doubt used for effect

to stimulate ye olde brainpan.

I was refereing to this from Fred's previous message:

for

the

4th harmonic but not the 4th harmonic, or some tendency to chaotically

into a different mode. The previous people cannot be complete

these kinds of issues are widely none thanks to authors

so there must be something up with this particular crystal

reliable design particularly challenging. If he doesn't

physics on this one, his product will be no more

Wow! I'm shocked. Fred actually made a funny typo. ;-)

--
Service to my country?
Been there, Done that, and I\'ve got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

Jim Thompson schrieb:

Maybe it helps to mount one (or a couple of them one after the other for statistics) into an appropriate fixture of a network analyzer to see which modes are excited and which not. The result should reflect all the crystal mechanics in the electrical amplitude, frequency and phase domain. It then should be possible to disentangle the results into values of a reasonable electrical model. Is this naive? I have never done it myself but it makes me eager to start tinkering with it too. Has anyone experience with this? Regards Gerhart